Learning: rapid aversive conditioning in the gastropod mollusk Pleurobranchaea.

Untrained Pleurobranchaea feed voraciously when presented food and withdraw from electrical shocks. We trained experimental animals in ten trials spaced 1 hour apart to withdraw from food alone by electrically shocking them if they fed or were indifferent to food. The greatest increase in the number of learned withdrawal responses occurred within 12 hours after conditioning, and was accompanied by long-lasting increased in the threshold and latency of feeding responses. Control animals, which received food and shock alternately (unpaired) every half hour, showed considerably weaker changes than experimentals. These control responses quickly returned to initial levels after conditioning.

Learning: classical and avoidance conditioning the mollusk Pleurobranchaea.

Naive specimens of the marine gastropod Pleurobranchaea withdraw from tactile stimulation of the oral veil and show feeding responses to food chemicals. Experimental subjects, trained by pairing touch (conditioned stimullus) with food chemicals (unconditioned stimulus), soon acquired a classically conditioned feeding response to touch alone. Control subjects that received touch alone or unpaired touch and food chemicals showed significantly fewer feeding responses to touch than did experimentals. Classically conditioned specimens were used for avoidance conditioning. Subjects that received aversive electrical stimulation when they did not withdraw from touch rapidly learned to withdraw rather than to feed in response to touch alone. Controls that received touch alone or unpaired touch and shock continued to exhibit the feeding response to touch alone. The learned responses persisted for up to 2 weeks without reinforcement before extinction, and could be demonstrated in the isolated nervous system.

Comparative study of histamine immunoreactivity in nervous systems of Aplysia and Pleurobranchaea.

The distribution of histamine in the nervous system of the marine molluscs Aplysia californica and Pleurobranchaea californica was studied by using a newly available immunohistochemical localization technique and specific antiserum against histamine-protein conjugate. We examined several sets of complete histological sections through the major ganglia of both animals, as well as all nerve roots of the buccal and cerebral ganglia and the corresponding target tissues. The results indicate that histamine is present in several neurons and/or nerve fibers of all major ganglia. An especially dense histamine fiber network in the buccal ganglion of both species suggests a major role for histamine in regulation of buccal-oral behaviors. Histamine was also observed in several identified nerve roots of the buccal and cerebral ganglia, as well as in the corresponding target tissues. Its localization in the Aplysia radular sac and in the statocyst neurons of both species suggest a role for histamine in sensory functions. Our study revealed many previously unknown histamine cells or cell clusters, some of which may be identifiable by electrophysiological methods. The findings also point to possible reinterpretation of previous findings, indicating that histamine may be a cotransmitter in identified cells, whereas the methodology itself suggests that special precautions must be taken to avoid spurious interpretations of specificity. As has now been observed in studies of serotonergic immunohistochemistry and in our own findings on VIP, histamine terminals were observed to lie in close contact with somata and axon hillocks, all of which suggest that axo-somatic connections in molluscs may be more prevalent than previously considered.

Immunohistochemistry of enkephalins: model studies on hapten-carrier conjugates and fixation methods.

For immunohistochemical demonstration of the enkephalin octapeptide Met5-enkephalin-Arg6-Gly7-Leu8, the peptide was conjugated with a carrier protein using either glutaraldehyde or 1-ethyl-3 (3-dimethylaminopropyl)-carbodiimide as coupling agent. Antisera were raised in rabbits and their specificity was studied using the immunoblotting technique. The results suggest that glutaraldehyde selectively couples the amino terminus of the peptide to the carrier protein, while carbodiimide coupling produces a mixture of specificities. Accordingly, antiserum raised against the glutaraldehyde-induced conjugate specifically recognized the peptide carboxyl terminus and allowed immunohistochemical distinction of the octapeptide from other closely related opioid peptides, such as Leu5- and Met5-enkephalin, Met5-enkephalin-Arg6-Phe7, and Phe1-Met2-Arg3-Phe4-NH2. In contrast, antiserum raised against the carbodiimide-induced octapeptide conjugate showed a mixture of specificities. Addition of glutaraldehyde to the fixative enhanced octapeptide immunoreactivity in several tissues and revealed a previously unknown nerve system in the pituitary gland. These results support the idea that optimal immunohistochemical demonstration of small molecules, which requires conjugation to a carrier protein, is obtained when the coupling agent is included in the fixative so as to induce the actual coupling reaction during fixation.

Putative neuroendocrine cells in the Aplysia cerebral ganglion.

We report on a special population of cells in the Aplysia cerebral ganglion that are characterized by several features compatible with neuroendocrine function. These cells can be recognized in living ganglia by their small size, white color and their typical distribution as a compact cluster in the central medial region of the dorsal ganglion surface. Upon intracellular recording, these cells generate action potentials of relatively long duration (about 25 ms), as compared with the faster action potentials of larger white cells or of non-white cells (about 4 ms). Intracellular injection of the small white cells with Lucifer yellow after recording revealed a dual projection area: single cells have one process which branches extensively into many varicose terminals as it courses through the neuropil, and then sends varicose terminals to the vascular sheath at the periphery of the ganglion. In cryostat sections, these cells were specifically characterized by their content of large granules, the staining characteristics of which distinguish them from lipochondria or lysosomes. Their ability to bind fluorochromes nonspecifically is of particular importance for the interpretation of histochemical localization studies based on immunofluorescence techniques.

Distribution of Met5-enkephalin-Arg6-Gly7-Leu8-immunoreactivity in the rat and mouse pituitary gland.

The distribution of the octapeptide Met5-enkephalin-Arg6-Gly7-Leu8 (MEAGL), a proenkephalin A-derived opioid peptide, in the rat and mouse pituitary gland was studied using the indirect immunofluorescence technique and immunoelectron microscopy. The anterior lobe contained a few MEAGL-immunoreactive cells but no nerve fibers. A previously unknown enkephalin-immunoreactive nerve fiber system was revealed in the intermediate lobe. These fibers originated in a dense MEAGL-immunoreactive plexus located along the border between the intermediate and posterior lobes and were distributed throughout the lobe. In the posterior lobe, MEAGL immunoreactivity was found in a very dense network of varicose fibers that was evenly distributed over the entire lobe. These results provide a morphological correlate for previous chemical studies and together with them suggest that MEAGL-immunoreactive innervation regulates endocrine functions of the intermediate and posterior lobes directly at the pituitary level.

Evidence for heterogeneity of muscarinic receptors in the mollusc Pleurobranchaea.

The properties of the specific binding of the muscarinic antagonist [125I]3-quinuclidinyl-4-iodobenzilate ([125I]4IQNB] to nervous tissue of Pleurobranchaea california were characterized. The specific binding of [125I]4IQNB to Pleurobranchaea nervous tissue was characterized by its high affinity (Kd = 0.61 +/- 0.11 nM) and saturability (Bmax = 602 +/- 46 fmol/mg protein). A comparison of the numbers of binding sites recognized by [125I]4IQNB and l-[3H]QNB in nervous tissue of three invertebrate species indicated that in Aplysia and Cancer magister (crab) ganglia membranes the two radioligands labeled comparable numbers of binding sites; however, in Pleurobranchaea membranes l-[3H]QNB recognized only a subpopulation (8-10%) of the total number of [125I]4IQNB binding sites. The disparity in the numbers of binding sites labeled by these radioligands was consistent with our finding of a heterogeneity of muscarinic antagonist binding sites in l-QNB competition experiments in Pleurobranchaea. Computer-assisted analysis of l-QNB competition of [125I]4IQNB specific binding demonstrated that these data were best described by a two-site model with high- and low-affinity sites for l-QNB. The high-affinity site recognized by l-QNB possessed an IC50 value of 0.2 nM and comprised 18% of the total specific binding, while the lower affinity site had an IC50 value of 55.6 nM and comprised the remaining 82% of the total population of [125I]4IQNB recognition sites. The IC50 value for l-QNB at the high-affinity site in Pleurobranchaea membranes is in excellent agreement with Kd values for l-[3H]QNB labeling of classical muscarinic receptors in a variety of invertebrate and vertebrate species.

Connectionist networks learn to transmit chaos.

Evidence presented in the preceding paper indicates that the activity of some neurons during the generation of coordinated motor patterns may be attributable to chaos. Because even "simple" biological systems are difficult to control, we have used connectionist networks in order to inquire into the question of whether a chaotic signal originating in one part of the nervous system can be learned and transmitted by another. We have examined a number of different architectures, and report here the findings for a simple network consisting of one input unit, four hidden units, and one output unit. During training sessions, the input of the circuit was given analog values of either the 3.60 or 3.95 logistic equation, or of one variable of the three-variable Rössler attractor. The backpropagated error in the learning algorithm was a function of the difference between the input value and the output at each iteration. Iterations involving small changes in analog value resulted in good similarity between the input and output signals, but little learning occurred because of the small error propagated back to the synapses. With larger differences in the analog values (and larger feedback error) at each iteration, we found that networks learned to transmit different chaotic attractors. Once the network learned one input, it could transmit another without changing the synapses. Increasing the number of hidden units increased the rate of learning.(ABSTRACT TRUNCATED AT 250 WORDS)

Muscarinic antagonist enhances one-trial food-aversion learning in the mollusc Pleurobranchaea.

One hour before training, in two replicate studies on the sea slug Pleurobranchaea californica, all animals (N = 114) received body-cavity injections of scopolamine, oxotremorine, or the equivalent volume of the saline/seawater vehicle that was used to inject the drugs. The low drug doses (2 mumol/kg) were near the threshold for generating observable neurophysiological responses, but did not affect feeding thresholds arising to a stimulus derived from beer (Sbr) and to one derived from squid (Ssq). Before training, the animals did not discriminate between Sbr and Ssq, as indicated by similar thresholds to both stimuli. During training, experimental animals in each injection group received Sbr alone for 10 sec and then paired with electric shocks for 50 sec; control animals received shocks 1 hr after Sbr. Postconditioning tests began 12 hr after training and were repeated once daily thereafter. After training, all experimental groups exhibited the expected aversive behavior to Sbr, as indicated by 100- to 1000-fold increases in feeding thresholds, and retained low thresholds to Ssq, but the scopolamine animals were better able to discriminate between Sbr and Ssq than either of the other experimental groups. The aversive responses to Sbr increased over a 3-day period, but there appeared to be no difference between injection groups on such a long-duration "consolidation" phase. Of the control groups, only the scopolamine animals exhibited low feeding thresholds to both Sbr and Ssq. The other control groups exhibited similar behavior as the experimental animals, indicating that associative factors relating to the 1-hr separation between Sbr and shocks may have produced the behavior in these control animals. Thus, by comparison to the other injections, scopolamine 1) increased the ability of the experimental animals to make the discrimination between Sbr and Ssq, and 2) prevented learning to avoid Sbr in the control animals. An accompanying paper provides a detailed characterization of muscarinic receptor pharmacology in Pleurobranchaea.

Evidence for chaos in spike trains of neurons that generate rhythmic motor patterns.

The findings presented here of work on the opisthobranch mollusc Pleurobranchaea californica indicate that some of the variability that has been observed in the activity of neurons during patterned motor activity may be attributable to low-dimensional chaos. We obtained long trains of action potentials (spikes) from these neurons, scanned them using adjacent temporal windows having equal widths, and converted the counts into frequency time series. These series were passed through a low-pass filter and detrended when necessary. The resulting time series gave a view of the envelopes of high-frequency bursts of spikes relating to the repetitive motor activity rather than of the intervals between spikes. Where applicable, we also compared analyses of smoothed data with the unprocessed spike intervals and found similar results for each type of time series. Autocorrelation functions of the processed data quickly decreased to zero, indicating that the long-term evolution of the time series could not be predicted from information at some given time. The first zero crossing of the autocorrelation function was used to define the lag for mapping the series into multidimensional phase space. These constructions were then used to examine the dynamics of the motor patterns directly from the state parameters of the time series: 1-D maps obtained from Poincaré slices of 2-D phase portraits, principal Lyapunov exponents, and correlation dimensions all indicated that the activity may be attributable to low-dimensional chaos. The present findings are similar to those of previous work in which equal-interval time series were obtained by interpolation of the unequal-interval spike trains. We discuss the implications of chaos and the difficulties in the application of extant dynamical tools to spike trains. An accompanying paper inquires into the ability of neural networks to read and transmit chaotic activity.

Attractors: architects of network organization?

An attractor is defined here informally as a state of activity toward which a system settles. The settling or relaxation process dissipates the effects produced by external perturbations. In neural systems the relaxation process occurs temporally in the responses of each neuron and spatially across the network such that the activity settles into a subset of the available connections. Within limits, the set of neurons toward which the coordinated neural firing settles can be different from one time to another, and a given set of neurons can generate different types of attractor activity, depending on how the input environment activates the network. Findings such as these indicate that though information resides in the details of neuroanatomic structure, the expression of this information is in the dynamics of attractors. As such, attractors are sources of information that can be used not only in adaptive behavior, but also to effect the neural architecture that generates the attractor. The discussion here focuses on the latter possibility. A conjecture is offered to show that the relaxation dynamic of an attractor may 'guide' activity-dependent learning processes in such a way that synaptic strengths, firing thresholds, the physical connections between neurons, and the size of the network are automatically set in an optimal, interrelated fashion. This inter-relatedness among network parameters would not be expected from more classical, 'switchboard' approaches to neural integration. The ideas are discussed within the context of 'pulse-propagated networks' or equivalently as 'spike-activated networks' in which the specific order in time intervals between action potentials carries important information for cooperative activity to emerge among neurons in a network. Though the proposed ideas are forward-looking, being based on preliminary work in biological and artificial networks, they are testable in biological neural networks reconstructed from identified neurons in cell culture and in simulation models of them.

Immunohistochemistry of Diverging and Converging Neurotransmitter Systems in Mollusks.

This series of studies was undertaken to compare the distribution of several transmitter-specific neuron systems in the nervous systems of the marine mollusks, Aplysia californica and Pleurobranchaea californica. Several specimens of each of the major ganglia of both species were sectioned serially, and each series was stained immunohistochemically to reveal one of the neuron systems. The present paper reports the results of stainings for acetylcholine, histamine, serotonin, gamma-aminobutyric acid (GABA), vasoactive intestinal polypeptide (VIP), cholecystokinin (CCK), Phe-Met-Arg-Phe-NH2 (FMRFamide), and small cardioactive peptide B (SCPB). For all the transmitter-specific sets of neurons examined, relatively few neurons send diverging projections to large areas of the neuropil in one or more ganglia. Moreover, different transmitters converge onto similar areas, and several transmitters evidently project onto the same identified neuron. Many of these diverging and converging projections are sufficiently extensive and overlapping that they are unlikely to be specific for a particular motorpattern. The overall findings are consistent with our previous neurophysiological data. These indicate that activity does not necessarily arise through circuit-specific, identifiable connections. Instead, appropriate response patterns, often containing mixtures of several behaviors, emerge variably through diffuse connections. These findings are also consistent with recent reports from other laboratories indicating that even light topical stimulation generates highly distributed neural activity, and that the responses of identifiable neurons are not constant. The anatomical data presented here should be combined with physiological experiments aimed at verifying the neurotransmitter action of the substances examined and elucidating the role of these agents in controlling the dynamics of neuronal responses.

The generation of rhythmic activity in a distributed motor system.

Rhythmic activity that is distributed to the brain and buccal ganglia and which underlies several types of behaviour, can be evoked from isolated nervous systems of Pleurobranchaea californica by tonic nerve stimulation. The experiments presented here were designed to test whether this rhythmic activity is produced by independent neuronal oscillators located in each ganglion or whether the rhythmic activity arises from a single oscillatory locus in the buccal ganglion and is transmitted passively to the brain. By interrupting the conduction of activity in the cerebrobuccal connectives (CBC) between brain and buccal ganglia we show that motor output from the brain depends on sustained, cycle to cycle input from the buccal ganglion and cannot be reset with respect to the buccal activity. The production of rhythmic activity in the brain depends on the generation of rhythmic activity in the buccal ganglia whether the rhythms are activated by stimulation of buccal roots or paracerebral command cells in the brain. Simultaneous intracellular recordings from brain motoneurones and buccal interneurones which project to the brain indicate that these interneurones provide both the drive and the pattern for rhythmic motor output in the brain. Tonic stimulation of the CBC can produce rhythmic activity in isolated brains in which all nerve roots and connectives have been cut. This can be explained by the fact that tonic stimulation of the connectives is transformed into phasic activity by the axons within the connective. We conclude therefore, that rhythmic, coordinated activity in the brain and buccal ganglia of Pleurobranchaea arises from oscillatory circuits that are located only in the buccal ganglia.

Discriminative behavior and Pavlovian conditioning in the mollusc Pleurobranchaea.

The buccal motor system in the sea slug Pleurobranchaea californica is multifunctional; similar sets of neurons and muscles generate different behaviors through similar electrophysiological motor patterns. Such multifunctional systems compromise the traditional practice of identifying a motor pattern and then using that pattern to indicate the behavior in reduced preparations. We address this issue in a series of experiments leading to the comparison of differential Pavlovian conditioning in whole animals with the conditioned behavior of the same animals during electrophysiological recording. Because differential conditioning requires two conditioned stimuli (CSs), we show here that each of two CSs activated the conditioned response from animals after they received the stimulus (CS+) paired with an unconditioned stimulus (UCS). Conditioning sessions consisted of 5 training trials with a 2-h intertrial interval. In one study, experimental animals received a 60-s CS+, derived from beer (Sbr), paired with a 50-s electrical shock UCS whose onset occurred 10 s after the CS+ onset; control animals received the Sbr and UCS explicitly unpaired. In a second study, animals received similar procedures as in the first but with a CS+ consisting of squid homogenate (Ssq). Tests with both CSs showed that animals did not discriminate between Sbr and Ssq before beginning conditioning, but did so afterward. Experimental animals exhibited robust food aversion (withdrawal and suppressed feeding) to the CS+, but retained strong appetitive responses to the CS they did not receive in training; response thresholds to the CS+ changed as much as 1000-fold by comparison to the preconditioning values. Control animals exhibited similar though significantly smaller behavioral changes as the experimental animals. Both stimuli evoked associatively learned responses, but Sbr produced greater experimental-control differences than Ssq did. Two accompanying papers show the results of using both CSs in differential conditioning, and describe the behavioral/electrophysiological comparisons.

Differential Pavlovian conditioning in the mollusc Pleurobranchaea.

The present differential Pavlovian conditioning experiments on the sea slug Pleurobranchaea californica extend conditioning described in a preceding paper and provide the conditioning foundation for studies reported in another accompanying paper comparing learned behavior in whole animals with the behavior and motor patterns of electrophysiological preparations. All animals received two appetitive-conditioned stimuli (CSs), one derived from beer (Sbr) and the other derived from squid muscle (Ssq), in different temporal relationships to an electric shock unconditioned stimulus (UCS). Two groups of animals were run concurrently. One group (n = 19) received Sbr as the CS+ in close temporal pairing with the UCS, and Ssq as the CS- explicitly unpaired with the UCS (Sbr +/Ssq-). The second group (n = 20) received the opposite contingencies (Sbr-/Ssq+). All animals received only one day of conditioning involving 5 trials with an intertrial interval of 2 h. There were two replicate experiments, each involving about half of the total n, and each yielding similar results as the sum we report here. Before conditioning, animals exhibited feeding behavior (extension of the proboscis and bite-strike responses) to both stimuli at similar low thresholds. Conditioning produced long-term behavioral changes in all animals throughout the 4.5-day postconditioning observation period. However, only the Sbr+/Ssq- animals consistently exhibited the appropriate differentially conditioned food-aversion behavior which consisted of strong withdrawal and high-threshold feeding responses to Sbr, and low-threshold feeding responses to Ssq. We discuss the possibility that such differences between Sbr+/Ssq- and Sbr-/Ssq+ conditioning may arise either from inherent differences in the responses of the animals to Sbr and Ssq, or, as seems more likely to us, from training and testing effects produced by differences in the compositions of the two stimuli.

Comparison of differential Pavlovian conditioning in whole animals and physiological preparations of Pleurobranchaea: implications of motor pattern variability.

The present study compares differential Pavlovian conditioning in whole animals with the behavior of the same animals during electrophysiological recording. Untrained specimens of the sea slug Pleurobranchaea did not discriminate between two appetitive stimuli, one derived from an extract of beer (Sbr) and the other from a homogenate of squid muscle (Ssq). When animals received Sbr as the CS+ and Ssq as the CS- in a single day of five-trial, differential Pavlovian conditioning they learned to avoid selectively the Sbr but continued to exhibit appetitive responses to Ssq. Quantitative measures show that there was over a 1000-fold increase in the thresholds of the proboscis extension and bite-strike responses, many animals ceased all feeding behavior, and exhibited withdrawal responses to Sbr. We examined the behavior of the same trained animals immediately before preparing them for physiological recording and during the recording session. There was a close one-to-one relationship between these behavioral observations, showing that the qualitative and quantitative features of whole-animal Pavlovian conditioning persist into the physiological preparations. Unexpectedly, motor patterns from untrained preparations showed considerable variability both within the same preparation at different times and between preparations; conditioning appeared to increase such variability. Thus, it was not possible to state unequivocally the behavior of the animal by examining the electromyogram recording alone. Many of the trained preparations not only exhibited suppressed feeding behavior and withdrawal responses to Sbr, but, as a consequence of the multifunctional nature of the Pleurobranchaea buccal-oral system, also regurgitated previously ingested Ssq or squid meat when they were stimulated with Sbr. We discuss the findings with respect to self-organizing mechanisms that may establish motor patterns in multifunctional systems, and suggest that such mechanisms may lead to the generation of behaviors that are not specifically encoded by the conditioned cellular changes.

Convergence in a distributed nervous system: parallel processing and self-organization.

The present findings show that the motor system of the carnivorous sea slug Pleurobranchaea californica consists of parallel, distributed, and interconnected neuronal channels by which motor activity may emerge from the dynamics of the system rather than from "switchboard" circuitry. The findings are shown primarily through the properties of the buccal-cerebral neurons (BCNs) that extensively converge and diverge monosynaptically and polysynaptically onto brain motoneurons, providing them with drive and patterned activity. The motoneurons, some of which are electrically coupled, feed back onto the BCNs. The BCNs are functionally heterogeneous both as a group and individually. Many are multifunctional in that they take part in the generation of different behaviors, and some also appear to change their timing with respect to the phase of the pattern generator in the different motor patterns. In the buccal ganglion, the BCNs affect the characteristics of the pattern generator and may be part of the pattern generator itself. By sending axons to buccal roots and to the brain, some BCNs may act as motoneurons and also integrate the activity of brain motoneurons. Because of the effects produced by the extensive interconnections among such functionally heterogeneous and nonlinear elements, and because the "history" of activity in the system can bias subsequent activity, there is ambiguity in assessing the response properties of neurons by examining them individually or in pairs. Such an assessment requires, first, an understanding of the context of activity in which a neuron becomes coactive, and, second, because of inherent variability in the system, it is necessary to consider the temporal, nonlinear computations of the system as a whole. We discuss the findings with regard to the attractor theory that has been used to study complex mammalian systems but that does not rely on modeling of any neuronal activity. The Pleurobranchaea nervous system may provide the means for studying individual neurons within such analyses of global activity.

Stereoselective L-[3H]quinuclidinyl benzilate-binding sites in nervous tissue of Aplysia californica: evidence for muscarinic receptors.

The muscarinic antagonist L-[3H]quinuclidinyl benzilate (L-[3H]QNB) binds with a high affinity (Kd = 0.77 nM) to a single population of specific sites (Bmax = 47 fmol/mg of protein) in nervous tissue of the gastropod mollusc, Aplysia. The specific L-[3H]QNB binding is displaced stereoselectively by the enantiomers of benzetimide, dexetimide, and levetimide. The pharmacologically active enantiomer, dexetimide, is more potent than levetimide as an inhibitor of L-[3H]QNB binding. Moreover, the muscarinic cholinergic ligands, scopolamine, atropine, oxotremorine, and pilocarpine are effective inhibitors of the specific L-[3H]QNB binding, whereas nicotinic receptor antagonists, decamethonium and d-tubocurarine, are considerably less effective. These pharmacological characteristics of the L-[3H]QNB-binding site provide evidence for classical muscarinic receptors in Aplysia nervous tissue. The physiological relevance of the dexetimide-displaceable L-[3H]QNB-binding site was supported by the demonstration of the sensitivity of the specific binding to thermal denaturation. Specific binding of L-[3H]QNB was also detected in nervous tissue of another marine gastropod, Pleurobranchaea californica. The characteristics of the Aplysia L-[3H]QNB-binding site are in accordance with studies of numerous vertebrate and invertebrate tissues indicating that the muscarinic cholinergic receptor site has been highly conserved through evolution.